Solved!
Go to SolutionJ
jpmorgan17
51 Posts
1
1028
August 21st, 2009 14:00
Celerra Block-Level Dedupe
I have read the Using Celerra Data Deduplication manual and looked at the Celerra Command Reference Manual in the fs_dedupe section and I do not see any mention of how to set the deduplication level at file versus block level. In the "Achieving Storage Efficiency through EMC Celerra Data Deduplication" mentions four types of Deduplication:
File-Level deduplication, fixed block deduplication, variable block deduplication, and compression.
I am assuming that the check-box and CLI turn on File-Level and compression, but how do you switch to block-level for greater space saving? Thanks!
File-Level deduplication, fixed block deduplication, variable block deduplication, and compression.
I am assuming that the check-box and CLI turn on File-Level and compression, but how do you switch to block-level for greater space saving? Thanks!
gbarretoxx1
2 Intern
•
366 Posts
364
0
Posted August 23rd, 2009 18:00
Hi,
Approximately 75 percent of the savings achieved by applying all four techniques above mentioned together can be achieved by applying compression alone. EMC has chosen a combination of compression and file-level deduplication technology to deliver the maximum storage savings with the least amount of resource footprint.
EMC has other products with other approaches that have the deduplication functionality on different levels.
Gustavo Barreto.jpmorgan17
51 Posts
364
0
Posted August 23rd, 2009 18:00
Thanks Gustavo! So does that mean EMC should not have released the white paper stating the other types of deduplication, or was it only to compare what Celerra currently does versus other technologies?Rainer_EMC
6 Operator
•
8645 Posts
364
1
Posted August 24th, 2009 04:00
Hi,
The white paper was to compare it to other technologies and to explain why single-instancing and file-level compression isnt necessarily worse than block-level deduplication.
Within a file for example compression is much more effizient than a fixed-length block-level dedupe - you can reduce much more than just one 4kB block and its not dependent on the blocks being the same at the same block boundaries.- i see compression being great for general office type of file server servers, i have a lot of shares with x-rays in jpeg format. I would probably get better space savings out of block level dedupe vs compression, jpeg is already compressed. What do you think ?
Rainer_EMC
6 Operator
•
8645 Posts
364
0
Posted August 24th, 2009 05:00
For already compressed files neither technique will give you any space savings.
compressed files are like encrypted files - they will generate different blocks for data even if its only slightly different.
gbarretoxx1
2 Intern
•
366 Posts
364
0
Posted August 22nd, 2009 22:00
Celerra currently supports only file-level deduplication.
Regards,
Gustavo Barreto.